[AMD Official Use Only - AMD Internal Distribution Only]
The series was
Reviewed-by: Likun Gao <Likun.Gao@xxxxxxx>.
Regards,
Likun
-----Original Message-----
From: Kenneth Feng <kenneth.feng@xxxxxxx>
Sent: Thursday, May 16, 2024 9:11 AM
To: amd-gfx@xxxxxxxxxxxxxxxxxxxxx
Cc: Gao, Likun <Likun.Gao@xxxxxxx>; Feng, Kenneth <Kenneth.Feng@xxxxxxx>
Subject: [PATCH 2/2] drm/amd/pm: enable thermal alert on smu 14.0.2/3
enable thermal alert on smu 14.0.2/3
Signed-off-by: Kenneth Feng <kenneth.feng@xxxxxxx>
---
drivers/gpu/drm/amd/pm/swsmu/inc/smu_v14_0.h | 8 +++
.../gpu/drm/amd/pm/swsmu/smu14/smu_v14_0.c | 68 ++++++++++++++++++-
.../drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c | 35 +++++++++-
3 files changed, 108 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v14_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v14_0.h
index 3d34b3869df6..9b97a4e95c0f 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v14_0.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v14_0.h
@@ -48,6 +48,10 @@
#define SMU14_TOOL_SIZE 0x19000
+#define CTF_OFFSET_EDGE 5
+#define CTF_OFFSET_HOTSPOT 5
+#define CTF_OFFSET_MEM 5
+
extern const int decoded_link_speed[5]; extern const int decoded_link_width[7];
@@ -236,5 +240,9 @@ int smu_v14_0_od_edit_dpm_table(struct smu_context *smu,
void smu_v14_0_set_smu_mailbox_registers(struct smu_context *smu);
+int smu_v14_0_enable_thermal_alert(struct smu_context *smu);
+
+int smu_v14_0_disable_thermal_alert(struct smu_context *smu);
+
#endif
#endif
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0.c
index fbfe9cae0e05..8cce17d1f230 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0.c
@@ -38,6 +38,8 @@
#include "amdgpu_ras.h"
#include "smu_cmn.h"
+#include "asic_reg/thm/thm_14_0_2_offset.h"
+#include "asic_reg/thm/thm_14_0_2_sh_mask.h"
#include "asic_reg/mp/mp_14_0_2_offset.h"
#include "asic_reg/mp/mp_14_0_2_sh_mask.h"
@@ -853,12 +855,19 @@ static int smu_v14_0_set_irq_state(struct amdgpu_device *adev,
unsigned tyep,
enum amdgpu_interrupt_state state) {
+ struct smu_context *smu = adev->powerplay.pp_handle;
+ uint32_t low, high;
uint32_t val = 0;
switch (state) {
case AMDGPU_IRQ_STATE_DISABLE:
/* For THM irqs */
- // TODO
+ val = RREG32_SOC15(THM, 0, regTHM_THERMAL_INT_CTRL);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, THERM_INTH_MASK, 1);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, THERM_INTL_MASK, 1);
+ WREG32_SOC15(THM, 0, regTHM_THERMAL_INT_CTRL, val);
+
+ WREG32_SOC15(THM, 0, regTHM_THERMAL_INT_ENA, 0);
/* For MP1 SW irqs */
if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(14, 0, 0) || @@ -875,7 +884,24 @@ static int smu_v14_0_set_irq_state(struct amdgpu_device *adev,
break;
case AMDGPU_IRQ_STATE_ENABLE:
/* For THM irqs */
- // TODO
+ low = max(SMU_THERMAL_MINIMUM_ALERT_TEMP,
+ smu->thermal_range.min / SMU_TEMPERATURE_UNITS_PER_CENTIGRADES);
+ high = min(SMU_THERMAL_MAXIMUM_ALERT_TEMP,
+ smu->thermal_range.software_shutdown_temp);
+ val = RREG32_SOC15(THM, 0, regTHM_THERMAL_INT_CTRL);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, MAX_IH_CREDIT, 5);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, THERM_IH_HW_ENA, 1);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, THERM_INTH_MASK, 0);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, THERM_INTL_MASK, 0);
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, DIG_THERM_INTH, (high & 0xff));
+ val = REG_SET_FIELD(val, THM_THERMAL_INT_CTRL, DIG_THERM_INTL, (low & 0xff));
+ val = val & (~THM_THERMAL_INT_CTRL__THERM_TRIGGER_MASK_MASK);
+ WREG32_SOC15(THM, 0, regTHM_THERMAL_INT_CTRL, val);
+
+ val = (1 << THM_THERMAL_INT_ENA__THERM_INTH_CLR__SHIFT);
+ val |= (1 << THM_THERMAL_INT_ENA__THERM_INTL_CLR__SHIFT);
+ val |= (1 << THM_THERMAL_INT_ENA__THERM_TRIGGER_CLR__SHIFT);
+ WREG32_SOC15(THM, 0, regTHM_THERMAL_INT_ENA, val);
/* For MP1 SW irqs */
if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(14, 0, 0) || @@ -1849,3 +1875,41 @@ int smu_v14_0_od_edit_dpm_table(struct smu_context *smu,
return ret;
}
+static int smu_v14_0_allow_ih_interrupt(struct smu_context *smu) {
+ return smu_cmn_send_smc_msg(smu,
+ SMU_MSG_AllowIHHostInterrupt,
+ NULL);
+}
+
+static int smu_v14_0_process_pending_interrupt(struct smu_context *smu)
+{
+ int ret = 0;
+
+ if (smu_cmn_feature_is_enabled(smu, SMU_FEATURE_ACDC_BIT))
+ ret = smu_v14_0_allow_ih_interrupt(smu);
+
+ return ret;
+}
+
+int smu_v14_0_enable_thermal_alert(struct smu_context *smu) {
+ int ret = 0;
+
+ if (!smu->irq_source.num_types)
+ return 0;
+
+ ret = amdgpu_irq_get(smu->adev, &smu->irq_source, 0);
+ if (ret)
+ return ret;
+
+ return smu_v14_0_process_pending_interrupt(smu);
+}
+
+int smu_v14_0_disable_thermal_alert(struct smu_context *smu) {
+ if (!smu->irq_source.num_types)
+ return 0;
+
+ return amdgpu_irq_put(smu->adev, &smu->irq_source, 0); }
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
index 1fbceb85d319..2b874e1ba4ea 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
@@ -1276,10 +1276,41 @@ static int smu_v14_0_2_update_pcie_parameters(struct smu_context *smu,
return 0;
}
+static const struct smu_temperature_range smu14_thermal_policy[] = {
+ {-273150, 99000, 99000, -273150, 99000, 99000, -273150, 99000, 99000},
+ { 120000, 120000, 120000, 120000, 120000, 120000, 120000, 120000,
+120000}, };
+
static int smu_v14_0_2_get_thermal_temperature_range(struct smu_context *smu,
struct smu_temperature_range *range) {
- // TODO
+ struct smu_table_context *table_context = &smu->smu_table;
+ struct smu_14_0_2_powerplay_table *powerplay_table =
+ table_context->power_play_table;
+ PPTable_t *pptable = smu->smu_table.driver_pptable;
+
+ if (amdgpu_sriov_vf(smu->adev))
+ return 0;
+
+ if (!range)
+ return -EINVAL;
+
+ memcpy(range, &smu14_thermal_policy[0], sizeof(struct
+smu_temperature_range));
+
+ range->max = pptable->CustomSkuTable.TemperatureLimit[TEMP_EDGE] *
+ SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
+ range->edge_emergency_max = (pptable->CustomSkuTable.TemperatureLimit[TEMP_EDGE] + CTF_OFFSET_EDGE) *
+ SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
+ range->hotspot_crit_max = pptable->CustomSkuTable.TemperatureLimit[TEMP_HOTSPOT] *
+ SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
+ range->hotspot_emergency_max = (pptable->CustomSkuTable.TemperatureLimit[TEMP_HOTSPOT] + CTF_OFFSET_HOTSPOT) *
+ SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
+ range->mem_crit_max = pptable->CustomSkuTable.TemperatureLimit[TEMP_MEM] *
+ SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
+ range->mem_emergency_max = (pptable->CustomSkuTable.TemperatureLimit[TEMP_MEM] + CTF_OFFSET_MEM)*
+ SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
+ range->software_shutdown_temp = powerplay_table->software_shutdown_temp;
+ range->software_shutdown_temp_offset =
+pptable->CustomSkuTable.FanAbnormalTempLimitOffset;
return 0;
}
@@ -1866,6 +1897,8 @@ static const struct pptable_funcs smu_v14_0_2_ppt_funcs = {
.update_pcie_parameters = smu_v14_0_2_update_pcie_parameters,
.get_thermal_temperature_range = smu_v14_0_2_get_thermal_temperature_range,
.register_irq_handler = smu_v14_0_register_irq_handler,
+ .enable_thermal_alert = smu_v14_0_enable_thermal_alert,
+ .disable_thermal_alert = smu_v14_0_disable_thermal_alert,
.notify_memory_pool_location = smu_v14_0_notify_memory_pool_location,
.set_soft_freq_limited_range = smu_v14_0_set_soft_freq_limited_range,
.init_pptable_microcode = smu_v14_0_init_pptable_microcode,
--
2.34.1
